Nutrient Limitation of Primary Productivity in the Southeast Pacific (BIOSOPE Cruise) Sophie Bonnet, C

Home , North Pacific Gyre, South Pacific Gyre

Nutrient limitation of primary productivity in the Southeast Pacific (BIOSOPE cruise) Sophie Bonnet, C. Guieu, F. Bruyant, O. Prášil, France van Wambeke, Patrick Raimbault, T. Moutin, C. Grob, M. Y. Gorbunov, J. P. Zehr, et al.

To cite this version:

Sophie Bonnet, C. Guieu, F. Bruyant, O. Prášil, France van Wambeke, et al.. Nutrient limitation of primary productivity in the Southeast Pacific (BIOSOPE cruise). Biogeosciences, European Geo- sciences Union, 2008, 5 (1), pp.215-225. hal-00330343

HAL Id: hal-00330343 https://hal.archives-ouvertes.fr/hal-00330343 Submitted on 14 Oct 2008

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Biogeosciences, 5, 215–225, 2008 www.biogeosciences.net/5/215/2008/ Biogeosciences © Author(s) 2008. This work is licensed under a Creative Commons License.

Nutrient limitation of primary productivity in the Southeast Paciﬁc (BIOSOPE cruise)

S. Bonnet1, C. Guieu1, F. Bruyant2, O. Pra´silˇ 3, F. Van Wambeke4, P. Raimbault4, T. Moutin4, C. Grob1, M. Y. Gorbunov5, J. P. Zehr6, S. M. Masquelier7, L. Garczarek7, and H. Claustre1 1Laboratoire d’Oceanographie´ de Villefranche, UMR 7093, CNRS and Universite´ Pierre et Marie Curie, BP 08 06238 Villefranche sur mer Cedex, France 2Dalhousie University–Department of Oceanography, 1355 Oxford Street Halifax, NS, B3H 4J1, Canada 3Institute of Microbiology ASCR, Opatovicky´ mlyn,´ 37981 Trebon and University of South Bohemia, Zamek,´ 37333 Nove´ Hrady, Czech Republic 4Centre d’Oceanologie´ de Marseille - Campus de Luminy, 13288 Marseille Cedex 09, France 5Rutgers University, Institute of marine and Costal Sciences, 71 Dudley road, New Brunswick, N. J. 08901-8521, USA 6University of California Santa Cruz, Ocean Sciences Department, 1156 high street, Santa Cruz, CA 95064, USA 7Station Biologique de Roscoff, UMR 7144, CNRS and Univ. Pierre et Marie Curie, BP 74, 29682 Roscoff Cedex, France Received: 2 July 2007 – Published in Biogeosciences Discuss.: 9 August 2007 Revised: 19 November 2007 – Accepted: 11 January 2008 – Published: 20 February 2008

Abstract. Iron is an essential nutrient involved in a variety 1 Introduction of biological processes in the ocean, including photosynthesis, respiration and dinitrogen fixation. Atmospheric depo- The production of organic matter in the sea is sustained by sition of aerosols is recognized as the main source of iron a continuous supply of essential macro- (C, N, P) and mi- for the surface ocean. In high nutrient, low chlorophyll ar- cronutrients (metals, vitamins). The nutrients requirements eas, it is now clearly established that iron limits phytoplank- vary among different phytoplanktonic species. According to ton productivity but its biogeochemical role in low nutrient, the Liebig’s law, organic matter production is controlled by low chlorophyll environments has been poorly studied. We the element that is available in the lowest concentration rel- investigated this question in the unexplored southeast Pa- ative to the growth requirements. This simple view is now cific, arguably the most oligotrophic area of the global ocean. replaced by the realization that multiple resources simultane- Situated far from any continental aerosol source, the atmo- ously limit phytoplankton growth in some parts of the ocean spheric iron flux to this province is amongst the lowest of (Arrigo, 2005). Global environmental forcings, including the world ocean. Here we report that, despite low dissolved human-induced climate change, could potentially modify the −1 iron concentrations (∼0.1 nmol l ) across the whole gyre nutrient delivery processes to the ocean, leading to funda- (3 stations located in the center and at the western and the mental changes in the diversity and functioning of the ma- eastern edges), primary productivity are only limited by iron rine food web. It is thus fundamental to understand which availability at the border of the gyre, but not in the center. nutrients control primary productivity in the open ocean to The seasonal stability of the gyre has apparently allowed for predict the biogeochemical consequences of global change. the development of populations acclimated to these extreme Representing 60% of the global ocean’s area, the subtropi- oligotrophic conditions. Moreover, despite clear evidence cal open-ocean ecosystems are the largest coherent biomes of nitrogen limitation in the central gyre, we were unable on our planet, and the biogeochemical processes they sup- to measure dinitrogen fixation in our experiments, even af- port are of global importance (Emerson et al., 1997; Karl, ter iron and/or phosphate additions, and cyanobacterial nif H 2002). The development of permanent time series stations gene abundances were extremely low compared to the North in the North tropical Atlantic and Pacific over the past two Pacific Gyre. The South Pacific gyre is therefore unique with decades have led to a revolution in the understanding of the respect to the physiological status of its phytoplankton pop- mechanisms and controls of nutrient dynamics in these re- ulations. mote environments. These oceanic gyres provide ideal eco- logical niche for the development of nitrogen-fixing organisms (e.g. Karl et al., 2002). In the North subtropical and Correspondence to: S. Bonnet tropical Atlantic and Pacific oceans, it has been estimated ([email protected]) that dinitrogen fixation is equivalent to 50–180% of the flux

Published by Copernicus Publications on behalf of the European Geosciences Union. 216 S. Bonnet et al.: The nutritional status of the South East Paciﬁc

a organisms have been suggested to be Fe-limited (Falkowski et al., 1998; Berman-Frank et al., 2001; Moore et al., 2002),

1 but direct experiment were lacking. 2 3 4 6 The BIOSOPE cruise provided the first spatially extensive 7 experiment in the Southeast Pacific. We conducted nutri- 12 13 14 15 ent addition bioassays, designed to investigate which nutrient (N, P and/or Fe) controls primary productivity, photosynthetic efficiency and dinitrogen fixation along a trophic gradient in the Southeast Pacific. A complementary paper (Van b Wambeke et al. this issue) examines the factors that control heterotrophic bacterial growth in the same area.

2 Material and methods

Fig. 1. (a) Transect of the BIOSOPE cruise from the Marquesas Is- This research was carried out onboard the R/V Atalante in lands to Chile superimposed on a SeaWiFS surface Chl-a compos- October–November 2004. The experiments were performed ite image (November–December 2004), and location of the short at three stations (Fig. 1a) located in the western edge (sta- (numbers) and long-stations of the cruise (MAR, HNL, GYR, EGY, tion HNL, 9◦ S 136◦ W), in the center (station GYR, 26◦ S UPW, UPX). This study reports results of bioassay experiments per- 114◦ W) and in the southeastern edge of the gyre (station formed at stations HNL, GYR and EGY. (b) Dissolved iron concen- EGY, 34◦ S 92◦ W). trations (0–400 m) (nmol l−1) along the BIOSOPE transect from Marquesas Islands (left) to Chile (right). All experimental setups were performed under strict trace metal clean conditions inside a clean container. Seawater was collected at 30 m depth using a trace metal-clean Teflon pump system and dispensed into 21 to 33 (depending on the − of NO3 into the euphotic zone (Karl et al., 1997; Capone number of treatments) acid-washed 4.5 l transparent polycar- et al., 2005), demonstrating that a large part of new primary bonate bottles. Under a laminar flow hood, nutrients or dust − productivity is fuelled by N2 fixation, rather than NO3 dif- were added either alone or in combination: +Fe, +NPSi and fusing from deeper layer into the euphotic zone. Dinitrogen +FeNPSi at station HNL, +Fe, +N, +P, +FeN, +FeNP and fixation requires the iron-rich nitrogenase complex, there- +dust at stations GYR and EGY. The final concentrations fore N2-fixing organisms have high iron (Fe) requirements −1 + −1 − −1 were 1 µmol l NH4 , 2 µmol l NaNO3 , 0.3 µmol l compared to phytoplankton growing on nitrate or ammonium −1 −1 NaH2 PO4, 2 nmol l FeCl3 and 0.25 mg l of dust. De- (Raven, 1988; Kustka et al., 2003). Dinitrogen fixation in spite the fact that Saharan dust deposition events are unlikely oceanic gyres has been seen to be controlled by Fe avail- to occur in the Southeast Pacific, the dust used in this exper- ability, as well as phosphate, which can (co)limit the process iment was the Saharan soils collected and characterized by (Sanudo-Wilhelmy et al., 2001; Mills et al., 2004). However, Guieu et al. (2002) in order to allow a comparison with ear- all studies dedicated to the nutrient control of primary pro- lier efforts (Mills et al., 2004; Blain et al., 2004, Bonnet et ductivity and dinitrogen fixation have focused so far on the al., 2005). Each fertilization was performed in triplicate. The Northern Hemisphere and there are few data available in the bottles were immediately capped with parafilm, sealed with Southern Hemisphere. PVC tape, and incubated for 48 h in an on-deck incubator The South Pacific gyre, which is the largest oceanic gyre of with circulating surface seawater at appropriated irradiance the global ocean, had been particularly undersampled (Claus- (50% ambient light level). For each station, the incubation tre et al. see introduction of this issue). This unique environ- started in the morning. After two selected time points during ment appears from satellite imagery and ocean colour to have the course of the experiment (T1=24 h; T2=48 h), three repli- the lowest chlorophyll-a (Chl-a) concentrations of the global cates of each treatment were sacrificed in order to measure ocean and thus represents an end-member of oceanic hyper- the following parameters: variable fluorescence, Chl-a con- oligotrophic conditions (Claustre and Maritorena, 2003). In centrations, epifluorescence microscopy counts and flow cy- contrast to the oceanic gyres located in the Northern Hemi- tometry. For rate measurements such as primary productivity sphere, the South Pacific Gyre is far removed from any con- and dinitrogen fixation, subsamples of each 4.5 l microcosms tinental source (anthropogenic and natural desert aerosols) have been used after each time point (24 h and 48 h) for par- and receives amongst the lowest atmospheric Fe flux in the allel incubations (see below). For each parameter, treatment world (Wagener et al., 2008). Consequently, the phytoplank- means were compared using a one-way ANOVA and a Fisher ton community as a whole, and particularly nitrogen-fixing PLSD means comparison test.

Biogeosciences, 5, 215–225, 2008 www.biogeosciences.net/5/215/2008/ S. Bonnet et al.: The nutritional status of the South East Paciﬁc 217

2.1 Nutrient analysis glycin references every batch of 10–15 samples. The ac- curacy of our analytical system was also regularly verified Ambient nutrient concentrations have been measured at each using reference materials from the International Atomic En- of the three stations before the incubation experiments as ergy Agency (AIEA, Analytical Quality Control Services). well as during the incubations. Concentrations of nitrate, Based on the lowest nitrogen level determined by our mass nitrite, phosphate and silicate have been analysed using a spectrometer (0.2 µmoles), the detection limit for dinitrogen Technicon Autoanalyser II (Treguer and Le Corre, 1975). fixation was 0.3 µmoles l−1. Dinitrogen fixation rates (rN in All the measurements have been done onboard, except for nmoles N l−1 t−1) were computed from an equation based the silicate samples, which were poisoned (mercuric chlo- on final particulate nitrogen (Dugdale and Wilkerson, 1986) ride 1 µg ml−1) and analyzed on land. Measurements in (see Raimbault and Garcia, this issue for more details on the the nanomolar range (lower detection limit=3 nmoles l−1) methodology). were obtained from the sensitive method described by Raim- bault et al. (1990). Nitrate at submicromolar levels (detec- 2.4 Primary productivity tion limit: 0.05 µmoles l−1) and phosphate (detection limit: 0.02 µ moles l−1) were measured according to Armstrong et al. (1967). Ammonium concentrations were measured as de- After each time point, primary production was measured scribed by Holmes et al. (1999), with a detection limit of on 250 ml subsamples of each microcosm as described by 5 nmoles l−1 Moutin and Raimbault (2002). Each subsample was inocu- 14 lated with 0.37 MBq of NaH CO3 (Amersham CFA3) and 2.2 Dissolved iron (DFe) concentrations incubated in a on-deck incubator (50% ambient light level) for four to seven hours around noon. Three samples were fil- DFe concentrations were measured on 19 vertical profiles tered immediately after inoculation for radioactivity determi- (0–400 m) along the 8000 km transect. They were analyzed nation at T0, and 250 µl were sampled randomly from three by Flow Injection Analysis with online preconcentration and bottles and stored with 250 µl of ethanolamine to determine chemiluminescence detection (FIA-CL) (adapted from Obata the quantity of added tracer (Qi). After incubation, the sam- et al., 1993). The mean blank, calculated from daily deter- ples were filtered on GF/F filters, covered with 500 µl of HCl minations, equaled 69±18 pmol l−1 (n=19) and the detection 0.5 mol l−1 and stored for pending analysis in the laboratory. limit was 54 pmol l−1. Each sample was analyzed in tripli- In the laboratory, samples were dried for 12 h at 60◦C, 10 ml cate. When at least two of the three runs agreed within ex- of ULTIMAGOLD-MV (Packard) were added to the filters pected reproducibility (10%), the average of the two or three and the radioactivity measured after 24 h with a Packard Tri concordant runs was taken as a correct concentration. If the carb 2100 TR liquid scintillation analyser. The hourly rate of concentration obtained deviated too much from the profile primary production (PP) was calculated as: −3 −1 continuum expectations and seemed to be contaminated, one PP(mgC m h )=(dpm-dpm(to))/(dpm(Qi)×1000)×25000/T of the other sampled bottles was then analyzed (in triplicate) with T=incubation duration. (see Blain et al. this issue, for more details on the methodology). 2.5 Flow cytometry 2.3 Dinitrogen fixation Cytometric analyses for picophytoplankton were performed At the last time point of each experiment, 0.6 ml of each mi- on fresh samples with a FACSCalibur (Becton Dickin- 15 crocosms have been subsampled. 1 ml of N2 gas (99% son) flow cytometer. Populations were differentiated based 15 N2 EURISOTOP) was introduced to each 0.6 l polycarbon- on their scattering and fluorescence signals (Marie et al., ate bottle through a Teflon-lined butyl rubber septum using a 2000). Samples were acquired for 3 minutes at ∼80 µl gas-tight syringe according to Montoya et al. (1996). After min−1 (∼12 to 100×103 cells) using Cell Quest Pro soft- 24 h of incubation, the samples were filtered under low vac- ware and data were analysed using the Cytowin software uum (100 mm Hg) through a precombusted (24 h at 450◦C) (see http://www.sb-roscoff.fr/Phyto/index.php?option=com 25-mm GF/F filter and dried at 60◦C. Filters were stored in docman\&task=cat view&gid=118\&Itemid=112). For- a desiccator until analysed. Determination of 15N enrich- ward scatter (FSC) and chlorophyll-a fluorescence (FL3) cy- ments was performed with an Integra-CN PDZ EUROPA tometric signals were normalized to reference beads (Flu- mass spectrometer. The background natural abundance de- oresbrite® YG Microspheres, Calibration Grade 1.00 µm, termined on 8 unlabelled samples was 0.367±0.007% for N. Polysciences, Inc) and then used as indicators of mean Only excess enrichments larger than twice the standard de- cell size and intracellular chlorophyll content, respectively viation (0.014% for N) were considered as significant. The (e.g. Campbell and Vaulot 1993). Significant changes in spectrometer was calibrated in order to detect the low levels mean FSC and FL3 after incubation under the different treat- of particulate nitrogen encountered; it was calibrated with ments (48 h) was evaluated through ANOVA analyses. www.biogeosciences.net/5/215/2008/ Biogeosciences, 5, 215–225, 2008 218 S. Bonnet et al.: The nutritional status of the South East Pacific

2.6 Epiﬂuorescence microscopy counts 3 Results

Counts were performed with a Olympus BX51 epifluo- 3.1 Initial nutrient concentrations and phytoplankton com- rescence microscope. Water samples (100 ml) were fixed position with glutaraldehyde (0.25% final concentration) and filtered through 0.8 µm pore size filters. Samples were stained with The initial nutrient concentrations and phytoplanktonic 4’6-diamidino-2-phenylindole (DAPI, 5 µg ml−1 final con- species composition for these bioassay experiments are given in Table 1. Fe vertical profiles indicated low centration). Eukaryotes were identified and counted by stan- −1 dard epifluorescence microscopy (Porter and Feig, 1980). (below 0.134±0.05 nmol l ) and constant DFe concentrations from the surface to 400 m depth throughout the entire transect (station MAR throughout sta- 2.7 Variable fluorescence tion EGY, n=110), except in the Chilean coastal upwelling zone (Fig. 1b, see also Blain et al. this is- Chlorophyll variable fluorescence of phytoplankton was sue). Surface DFe concentrations were 0.14±0.02 nmol l−1, measured using the custom-built benchtop Fluorescence 0.10±0.01 nmol l−1 and 0.10±0.01 nmol l−1, respectively Induction and Relaxation (FIRe) system (Gorbunov and for the HNL, GYR and EGY stations (Table 1). In Falkowski, 2005). The excitation light was provided by contrast, macronutrients and Chl-a concentrations differed 4 blue light-emitting diodes, LEDs, (central wavelength markedly among stations, with Chl-a concentrations be- −3 − 450 nm, 30 nm bandwidth, with the peak optical power den- ing 0.029±0.01 mg m in NO3 -depleted waters of GYR sity of 2 W cm−2). The variable fluorescence sequences were and 0.103±0.02 and 0.110±0.01 mg m−3 respectively at − processed to calculate minimum (Fo) and maximum (Fm) EGY and HNL, where NO3 concentrations were higher fluorescence (measured in the dark), as the quantum effi- (0.02±0.02 and 1.66±0.11 µmol l−1; Table 1). Waters were ciency of PSII (Fv/Fm), according to Kolber et al. (1998). phosphate-replete along the whole transect with concentra- Measurements were made on dark-adapted samples (30 min). tions always above 0.11 µmol l−1. The background fluorescence signal (blank) was measured The phytoplankton community structure was dominated using 0.2 µm filtered seawater and was subtracted from the by picophytoplankton at the three stations, where it repre- measured variable fluorescence. sented 58%, 49% and 47% of the total Chl-a (Table 1), respectively at HNL, GYR and EGY. Cyanobacteria, mainly 2.8 Abundance of nitrogen fixers belonging to the Prochlorococcus genera, dominated pico- phytoplancton. Phytoplankton pigment distribution along the Water samples (3 l) were filtered through 3 µm pore size fil- transect is described in Ras et al. (2007). ters (GE Osmonics) and subsequently through 0.2 µm pore size Supor filters (PALL corp.). Both filters were processed 3.2 Biological response during the incubation experiments to determine the N2-fixing microorganisms in the >3 µm and <3 µm size fractions. DNA was extracted from the For all the parameters measured in this experiments, the stim- filters (Church et al., 2005) with the addition of a bead- ulation by nutrients was considered to be significant when the beating step prior to the lysis step. The nif H gene was ANOVA comparison of distribution of triplicate treatments amplified with nested PCR primers (nif H1, nif H2, nif H3 gave values of p<0.05. The significant responses (different and nif H4) (Church et al., 2005). The amplification prod- from the control) are indicated by an asterisk on Fig. 2. ucts were cloned into pGEM®-T vectors (Promega). Plas- mid DNA was isolated with Montage kits (Millipore) and 3.2.1 Photosynthetic quantum efficiency of photosystem II the cloned inserts were sequenced at the University of The western part of the gyre (station HNL) was character- California-Berkeley Sequencing facility. Quantitative PCR ized by low F /F (0.16±0.01) at T0 (Table 1), indicating for Group A and B unicellular cyanobacterial nif H was per- v m an apparently low yield of photosynthesis. Iron was found formed as described in Church et al. (2005). to be the nutrient that controls photosynthetic efficiency at that station, as indicated by the increase of 65% of Fv/Fm 2.9 Determination of pigments after an Fe addition (p<0.05; Fig. 2). The station located at the eastern side of the gyre (station EGY) exhibited medium 2.8 l of seawater were filtered onto GF/F filters and imme- Fv/Fm values at T0 (0.30±0.02) (Table 1) that increased sig- diately stored in liquid nitrogen then at −80◦C until anal- nificantly (p<0.05) after Fe addition, but in a lower propor- ysis on land which was performed according to the proce- tion compared to station HNL (30%; Fig. 2). In contrast, the dure described in Ras et al. (2007). Pigment grouping into center of the South Pacific Gyre was characterized by high pigment-base size classes was performed according to Uitz Fv/Fm at T0 (0.51±0.03) (Table 1). This value didnot in- et al. (2006). crease after Fe, N, P or dust addition (p>0.05, Fig. 2). The

Biogeosciences, 5, 215–225, 2008 www.biogeosciences.net/5/215/2008/ S. Bonnet et al.: The nutritional status of the South East Paciﬁc 219

Fig. 2. Effect of nutrient additions during bioassay experiments performed at the three stations (HNL, GYR, EGY) (a) Photochemical −1 −1 efficiency of photosystem II (Fv/Fm) after 24 h incubation, (b) Primary productivity per unit chlorophyll-a (mg C mg Chl-a d ) after 48 h of incubation. Fv/Fm and carbon fixation were measured from separate triplicate bottles, such that nine bottles were incubated for each nutrient treatment. The error bars represent the standard deviation from triplicate incubations. Treatment means were compared using a one-way ANOVA and a Fisher PLSD means comparison test. Means that are significantly different from the control are labelled with an asterisk and means that are not significantly different are labelled with the same letter (p<0.05).

Table 1. Initial conditions for bioassay experiments. % pico-, nano- and microphytoplankton correspond to the biomass proportion of total Chl-a associated with each size class of phytoplankton calculated as described in Uitz et al. (2006).

Experiment 1 HNL Experiment 2 GYR Experiment 3 EGY Latitude Longitude 9◦04′S 136◦97′W 26◦04′S 114◦02′W 31◦89′S 91◦39′W Chl-a (mg m−3) 0.11±0.05 0.029±0.01 0.103±0.02 − −1 NO3 (µmol l ) 1.66±0.11 <0.003 0.02±0.02 − −1 NO2 (µmol l ) 0.03±0.00 < 0.003 0.01±0.00 + −1 NH4 (µmol l ) 0.03±0.00 0.006±0.00 < 0.005 3− −1 PO4 (µmol l ) 0.30±0.01 0.11±0.02 0.14±0.06 −1 SiOH4 (µmol l ) 1.31±0.12 0.98±0.09 1.21±0.17 DFe (nmol l−1) 0.14±0.02 0.10±0.01 0.10±0.01 Fv/Fm 0.16±0.01 0.51±0.03 0.30±0.01 % pico- 58 49 47 % nano- 31 48 43 % micro- 11 13 10

addition of macronutrients without Fe had a positive effect 3.2.2 Rate measurements at station HNL, but did not have any effect at station EGY, where only a Fe addition (alone or in combination) resulted in an increase of Fv/Fm. The dust treatment also had a posi- Primary productivity tive effect on Fv/Fm at the station EGY. At station HNL, only the addition of Fe resulted in a signiﬁcant increase of primary productivity (+50%, p<0.05), whereas at station GYR, only the treatments hav- ing nitrogen (+N, +FeN and +FeNP) resulted in a positive response (+45%, p<0.05). At station EGY, all treatments www.biogeosciences.net/5/215/2008/ Biogeosciences, 5, 215–225, 2008 220 S. Bonnet et al.: The nutritional status of the South East Paciﬁc

Table 2. Abundance of diazotrophs (gene copies per liter) determined by QPCR for samples collected at 13 stations between HNL and EGY (ND: Non Detectable). The depth at which the nif H gene was detectable is indicated between parentheses. The nitrogen ﬁxation rates (nmol l−1 d−1) measured at the three stations during the incubation experiments are also indicated. a Data obtained from Church et al. (2005) at 25 m depth in December 2002 (Aloha station). b Data from Dore et al. (2002).

< 3 µm > 3 µm

Station Group A Group B Trichodesmium Chaetoceros/Calothrix N2 ﬁxation rates HNL 27 (5 m) ND ND ND 1 ND ND ND ND 2 ND ND ND 30 (70 m) ND 3 ND ND ND ND 4 ND ND ND ND 6 – ND ND ND 7 ND ND ND ND GYR 184 (5 m) ND ND ND ND 12 – ND ND ND 13 ND ND ND ND 14 ND – – – 15 ND ND ND ND EGY ND ND ND ND ND ALOHA 10000–00000a 1000a 1000–10000a – 0.07–2.21b

with a source of Fe or N resulted in a positive increase the transect. It must be noted that the molecular data for (+25%). The addition of both Fe and N resulted in a higher both ends of the transect (data not related to these incuba- response (+50%), indicating a clear Fe and N co-limitation. tions) indicate the presence of larger number of cyanobacte- At stations GYR and EGY, the addition of P together with ria close to the Marquesas archipelago (up to 342 copies l−1 Fe and N did not result in a significantly higher response Group B, 45 copies l−1 Trichodesmium and 178 copies l−1 than the addition of Fe and N alone, indicating that P is not Chaetoceros/Calothrix) and close to the Chilean upwelling limiting. (108 copies l−1 Group B and 20 copies l−1 Trichodesmium). These higher densities of diazotrophs are consistent with the Dinitrogen fixation higher dinitrogen fixation rates measured by Raimbault et al. (2007) at both ends of the transect. 15 N2 assimilation consistently remained below the detection limit at the three stations in our incubation experiments 3.2.4 Cell numbers (Table 2), indicating the absence of dinitrogen fixation, even after dust, Fe, and FeP additions. Epifluorescence microscopy confirmed the absence of the two phylotypes Trichodesmium and unicellular from Group 3.2.3 Abundance of nitrogen fixers B at any station and treatment, which is in agreement with molecular biology data. Concerning non nitrogen-fixing or- Water samples from 13 stations situated between HNL and ganisms, the addition of Fe or Fe and/or macronutrients at EGY (Fig. 1a) were examined for presence of N2-fixing stations HNL resulted in a significant (p<0.05) increase in microorganisms by amplification of the nif H gene. Af- Synechococcus and picoeukaryotes abundances (Table 3). At ter amplification, cloning and sequencing the nitrogenase station EGY, all treatments containing nitrogen had a pos- genes, our results indicate the absence of the filamentous itive effect on picoeukaryotes, Prochlorococcus and Syne- cyanobacteria Trichodesmium, or any large (3–7 µm) uni- chococcus abundances. At station GYR, only the +N and cellular putative nitrogen fixing cyanobacteria (Group B). +Fe and N treatments induced a significant increase in pi- The results suggest however the presence of low numbers coeukaryotes abundances. The other treatments did not have of Group A cyanobacterial phylotypes at two stations (less any effect on this group, nor on Synechococcus. However, than 200 copies l−1, Table 2), and low numbers of non- larger Synechococcus fluorescence (FL3) and forward light cyanobacterial nif H sequences (Vibrio diazotrophicus and scatter (FSC) cytometric signals indicated at station GYR an proteobacteria), that must explain the significant dinitrogen increase in relative cell size and intracellular chlorophyll-a fixation rates measured by Raimbault et al. (2007) during content after addition of N, Fe and N and FeNP (p<0.05 for

Biogeosciences, 5, 215–225, 2008 www.biogeosciences.net/5/215/2008/ S. Bonnet et al.: The nutritional status of the South East Paciﬁc 221

Table 3. Evolution of the abundances of Prochlorococcus, Synechococcus and picophyto-eukaryotes after 48 h of incubation at the three stations studied, for each treatment (cells ml−1; mean± SD). The standard deviation (SD) is calculated on the triplicates. At station GYR, Prochlorococcus ﬂuorescence was too dim to allow us to detect changes in either the abundance or cytometric signals. Treatment means were compared using a one-way ANOVA and a Fisher PLSD means comparison test. Means that are signiﬁcantly different from the control (p<0.05) are labelled with an asterisk.

HNL station Sample Prochlorococcus/ml Synechococcus/ml Picoeukaryotes/ml Control 264133 ± 17091 30096 ± 677 10140 ± 371 Fe 116928 ± 1858 47740 ± 1256* 15502 ± 1039* NPSi 420368 ± 43935* 60892 ± 4157* 23309 ± 1056* FeNPSi 142137 ± 12420 61178 ± 1726* 18243 ± 1398*

GYR station Sample Prochlorococcus/ml Synechococcus/ml Picoeukaryotes/ml Control nd 1704 ± 272 572 ± 109 Fe nd 1873 ± 176 518 ± 37 N nd 1659 ± 119 775 ± 144* FeN nd 1491 ± 79 916 ± 106* FeNP nd 1546 ± 55 678 ± 113 Dust nd 2012 ± 129 625 ± 33

EGY station Sample Prochlorococcus/ml Synechococcus/ml Picoeukaryotes/ml Control 158798 ± 6284 18743 ± 512 6422 ± 162 Fe 102588 ± 3761 15963 ± 795 5724 ± 141 N 253317 ± 19927* 20521 ± 900* 8642 ± 589* FeN 129818 ± 12825 18129 ± 669 7960 ± 237* FeNP 127752 ± 7424 17400 ± 454 8694 ± 186* Dust 138278 ± nd 17936 ± 481 8288 ± 312*

Table 4. Cytometric signals obtained at station GYR for each treat- 4 Discussion ment. Significant changes in mean FSC and FL3 after incubation under the different treatments (48 h) was evaluated using a one- 4.1 The role of Fe way ANOVA. Significant data (p<0.05) are labelled with an asterisk. Abbreviations: Proc (Prochlorococcus), Syn (Synechococ- The data indicate that Fe is the nutrient that controls pho- cus), Euk (picophytoeukaryotes), SSC (side scattered light inten- tosynthetic efficiency and primary productivity outside the sity), FSC (Forward scattered light intensity), FL3 (chlorophyll-a gyre, at station HNL. These results are in accordance with fluorescence). the patterns found in other HNLC waters (e.g. Boyd et al., 2000). Around Marquesas, Behrenfeld and Kolber (1999) Fe N Fe and N All Dust also found low values of Fv/Fm with a pronounced decrease at night; in our experiments, the addition of Fe eliminated SSC Proc. 0.065* 0.013* 0.024* 0.034* 0.016* the nocturnal decrease and increased F /F values. In the FSC Syn. 0.470 0.012* 0.006* 0.009* 0.858 v m SSC Syn. 0.792 0.805 0.374 0.444 0.400 central gyre (GYR), the high Fv/Fm value (0.51±0.03) was FL3 Syn. 0.230 0.021* 0.020* 0.018* 0.708 however unexpected due to the low dissolved Fe concentra- FSC Euk. 0.646 0.011* 0.259 0.221 0.883 tions. This value is close to the maximum value observed in SSC Euk. 0.663 0.002* 0.124 0.048* 1.000 the ocean (Falkowski et al., 2004) and did not increase af- FL3 Euk. 0.096 0.032* 0.117 0.01* 0.595 ter Fe or dust addition. Behrenfeld et al. (2006) found the same pattern (high Fv/Fm, absence of nocturnal decrease) in the North Tropical Pacific, but the ambient dissolved Fe concentrations there are two to seven times higher (Boyle et al., 2005) than in the South Pacific Gyre, where our experiments FSC and FL3, Table 4). The presence of cyanobacteria in the were carried out. Our data also indicate that the addition of size range 1 to 3 µm was later confirmed by epifluorescence Fe did not change Chl-a concentrations or primary produc- microscopy counts. tivity (p>0.05), indicating that, in contrary to the HNL sta- www.biogeosciences.net/5/215/2008/ Biogeosciences, 5, 215–225, 2008 222 S. Bonnet et al.: The nutritional status of the South East Pacific tion, the photoautotrophic community was not Fe-limited in tion in station GYR. Station EGY, located on the southeast- the gyre. This suggests that the natural assemblage is accli- ern edge of the gyre, is a transition station where primary mated to Fe deprivation. productivity is Fe and N co-limited. In the center of the gyre Flow cytometry measurements identified Prochlorococcus (station GYR), nitrogen is the nutrient that controls primary as a prominent component of the prokaryote-dominated phy- productivity. The addition of a nitrogen source resulted in an toplankton assemblage at station GYR (20 000 cells ml−1), increase in the abundance of picophytoeukaryotes and an in- whereas in terms of carbon biomass, picophytoeukary- crease in relative cell size and intracellular Chl-a content of otes dominated (0.89 mg C m−3, i.e. 2.6-fold higher than Synechococcus (p<0.05). It is interesting to note that bac- Prochlorococcus). Although Synechococcus abundance was terial production is also directly enhanced after a nitrogen similar to that of Prochlorococcus (1400 cells ml−1), their addition (by a factor of 9 after 48 h, see Van Wambeke et al. contribution to the phytoplanktonic carbon biomass was neg- this issue). ligible (0.06 mg C m−3). To maintain high carbon fixation Perhaps the most intriguing part of this study is the ab- rates in such a low Fe environment, the organisms must sence of dinitrogen fixation in our experiments, even after have developed ecophysiological strategies to survive the dust, Fe and/or P additions. These results suggest that nei- shortage of Fe, including Fe scavenging systems (Geider ther P nor Fe limit dinitrogen fixation at 30 m depth. It and la Roche, 1994), efficient Fe transport systems over the must be noted that Raimbault and Garcia (2007) measured plasma membrane (Katoh et al., 2001) or gene regulation low dinitrogen fixation rates at 30 m in the central gyre (sta- systems consisting in rearrangements of photosynthetic ap- tions GYR), revealing the weakness of the process at this paratus (Sandstrom¨ et al., 2002). depth. However, these authors measured significant N2 fix- In summary, although DFe concentrations were identical ation rates in subsurface waters (∼1 nmol l−1 d−1), which is at the three stations, our data clearly show contrasting phys- in accordance with the molecular data (Table 2) and recent iological responses to Fe additions. Cultures experiments modelling efforts (Deutsch et al. 2007). At the other stations conducted under Fe limited conditions exhibit either low located between stations HNL and EGY, diazotrophic het- (∼0.1) or high (∼0.5) Fv/Fm depending on whether growth erotrophic bacteria detected by molecular tools must be re- is balanced or unbalanced (Price, 2005). The high Fv/Fm sponsible of the dinitrogen fixation rates measured at regular values measured in the center of the gyre (GYR) are a clear stations by Raimbault and Garcia (2007). It has to be noted indication that the phytoplankton assemblages are well ac- that the density of diazotrophic cyanobacteria are extremely climated to the stable environmental conditions of low N and low compared to those of the North Pacific Gyre (ALOHA low Fe. In contrast, station HNL (situated in the southern station in December, 10 000 to 100 000 copies l−1). The ab- limit of the equatorial upwelling and embedded in the west- sence of Trichodesmium and Group B phylotypes also con- ward flowing South Equatorial Current) and station EGY trasts with amplification from oligotrophic waters of the trop- (corresponding to a transition zone between the salty East- ical North Pacific Ocean, where cyanobacterial nif H genes ern South Pacific Central Waters and the waters influenced from these two groups are abundant, even during the winter by fresher Subantarctic Surface Waters (Emery and Meincke, season (see Table 2; Church et al., 2005). In this area, dini- 1986) are less steady environments, with low Fv/Fm values trogen fixation rates are high throughout the water column and increased Fo possibly due to the presence of specific Fe- (Dore et al., 2002) and provide a major source of newly fixed stress pigment-protein complexes (Behrenfeld et al., 2006). nitrogen to the euphotic zone, sustaining up to 50% of new This might also suggest an unbalanced growth (Parkhill et al., primary production, and drives the system towards P- and/or 2001) in environment with occasional spikes of nutrients. Fe-limitation (Karl et al., 1997; Sohm et al.1). The scarcity of It is interesting to note that a dust addition did not cause nitrogen fixing organisms in the South Pacific Gyre may be any increase of primary productivity at station EGY, while one of the origins of the relatively large phosphate concen- an iron addition resulted in a positive response. This absence trations (always above 0.11 µmol l−1; Moutin et al., 2007), of response can be interpreted by the fact that only 0.1% of as well as the N controlled status of the phytoplankton and the iron in the dust dissolved, which is ten times lower than bacterial communities. However, the other potential sources the dissolution found with the same amount of the same dust of nitrogen for the South Pacific Gyre are small and dinitro- in the Mediterranean waters by Bonnet and Guieu (2004). gen fixation may nonetheless represent the main source of This difference can be interpreted by the difference in organic new nitrogen in the system (Raimbault and Garcia, 2007): − ligands concentrations between the Pacific and the Mediter- the vertical flux of NO3 from below the thermocline is ex- ranean waters (Bonnet, 2005). tremely low compared to other gyres (10 to 12 times lower than the one measured in the North Atlantic gyre; Capone 4.2 From Fe to nitrogen limitation 1Sohm, J., Krauk, J., Mahaffey, C., Capone, D. G.: Diagnostics Dissolved Fe concentrations were low and constant along the of phosphorus stress in the cyanobacterium Trichodesmium reveal three stations studied, but the data clearly show a progression the northwest Atlantic is more severely P limited than the tropical from Fe limitation in station HNL towards nitrogen limita- Pacific, Limnol. Oceanogr., submitted, 2008.

Biogeosciences, 5, 215–225, 2008 www.biogeosciences.net/5/215/2008/ S. Bonnet et al.: The nutritional status of the South East Pacific 223 et al., 2005), and potential atmospheric deposition of nitro- References gen is almost absent according to the aerosols measurements performed in this area by Wagener et al. (2008). Armstrong, F. A. J., Stearns, C. R., and Strickland, J. D. H.: The measurement of upwelling and subsequent biological processes by means of the Technicon AutoAnalyzer and associated equip- 5 Conclusions ment, Deep Sea Res., 14(3), 381–389, 1967. Arrigo, K. R.: Marine micro-organisms and global nutrient cycles, This paper is an attempt to analyse which are the nutrients Nature, 437(7057), 349–355, 2005. that control primary productivity in the Southeast Pacific. Behrenfeld, M. J. and Kolber, Z. S.: Widespread iron limitation A gradient in the nutrient control of phytoplanktonic com- of phytoplankton in the South Pacific Ocean, Science, 238 840– munities is described, from an Fe-controlled system (on the 843, 1999. Behrenfeld, M. J., Worthington, K., Sherell, R. M., et al.: Controls edge of the gyre) towards a nitrogen-controlled system (in on tropical Pacific ocean productivity revealed through nutrient the center of the gyre). By combining physiological data, stress diagnostic, Nature 442, 1025–1028, 2006. rate processes measurements and molecular approaches, this Berman-Frank, I., J., Cullen, J., Hareli, Y., et al.: Iron availabil- unique nutritional status of the phytoplankton populations is ity, cellular iron quotas, and nitrogen fixation in Trichodesmium, described for the first time in the gyre. Autotrophic com- Limnol. Oceanogr., 46, 1249–1277, 2001. munities are indeed adapted to living under extremely low Blain, S., Bonnet, S., and Guieu, C.: Dissolved iron distribution Fe levels; these results add a new perspective on life in ex- in the tropical and sub tropical Southeastern Pacific, Biogeo- treme environments and give a new perspective to so-called sciences Discuss., 4, 2845–2875, 2007, HNLC areas of the world ocean. In addition, we show for the http://www.biogeosciences-discuss.net/4/2845/2007/. first time that nitrogen fixing cyanobacteria are scarce in the Bonnet, S. and Guieu, C.: Dissolution of atmospheric South Pacific Gyre. Due to the extremely low Fe-rich dust iron in seawater. Geophys. Res. Lett., 31, L03303, doi:10.1029/2003GL018423, 2004. inputs at the surface waters of the gyre, these waters con- Bonnet S., Guieu C., Chiaverini J., et al.: Impact of atmospheric institute a “low iron” environment (Blain et al. this issue). It puts on the autotrophic communities in a low nutrient low chloro- is hypothesized that this region is not a favourable environ- phyll system, Limnol. Oceanogr., 50(6), 1810–1819, 2005. ment for common photoautotrophic nitrogen fixing organ- Bonnet, S.: The role of atmospheric iron in oligotrophic environ- isms (e.g. Trichodesmium), as they have elevated Fe quo- ments. Ph.D. thesis, University Pierre et Marie Curie, Paris 6, tas relative to non-diazotrophic phytoplankton (Kustka et al., 207 pp., 2005. 2003). However, the factors controlling dinitrogen fixation Boyd, P. W., Watson, A. J., Law, C. S., et al.: A mesoscale phyto- are still poorly understood and further studies are needed to plankton bloom in the polar Southern Ocean stimulated by iron understand the distribution of these organisms and their bio- fertilization, Nature, 407, 695–702, 2000. geochemical impact in the ocean. Boyle, E., Bergquist, B. A., Kayser, R. A., and Mahowald, N.: Iron, manganese, and Lead at Hawaii Ocean Time-series sta- Acknowledgements. Dominique Tailliez and Claudie Bournot are tion ALOHA: Temporal variability and an intermediate water warmly thanked for their efficient help in CTD rosette management hydrothermal plume, Geochim. Cosmochim. Act., 69, 933–952, and data processing. This is a contribution of the BIOSOPE 2005. project of the LEFE-CYBER program. This research was funded Campbell, L. and Vaulot, D.: Photosynthetic picoplankton commu- by the Centre National de la Recherche Scientifique (CNRS), the nity structure in the subtropical North Pacific Ocean near Hawaii Institut des Sciences de l’Univers (INSU), the Centre National (station ALOHA).Deep-Sea Res., 40, 2043–2060, 1993. d’Etudes´ Spatiales (CNES), the European Space Agency (ESA), Capone, D. G., Burns, J. A., Montoya, J. P. et al.: New nitrogen The National Aeronautics and Space Administration (NASA), input to the tropical North Atlantic Ocean by nitrogen fixation the Academy of Sciences (AVCR) and the Ministry of Education by the cyanobacterium, Trichodesmium spp., Global Biogeoch. (MSMT) of the Czech Republic, and the Natural Sciences and Cy., 19, GB2024, doi:10.1029/2004GB002331, 2005. Engineering Research Council of Canada (NSERC). Church, M. J., Jenkins, B. D., Karl, D. M., and Zehr, J. P.: Vertical distributions of nitrogen-fixing phylotypes at Stn ALOHA in the Edited by: J.-P. Gattuso oligotrophic North Pacific Ocean. Aquat. Microb. Ecol., 38(1), 3–14, 2005. Claustre, H. and Maritorena, S.: The many shades of ocean blue, Science, 302, 1514–1515, 2003. Deutsch, C., Sarmiento, J. L., Sigman, D. M., et al.: Spatial cou- pling of nitrogen inputs and losses in the ocean, Nature, 445, 163–167, 2007. Dore, J. E., Brum, J. R., Tupas, L. M., and Karl, D. M.: Sea- sonal and Interannual Variability in Sources of Nitrogen Support- ing Export in the Oligotrophic Subtropical North Pacific Ocean. Limnol. Oceanogr, 47, 1595–1607, 2002. Dugdale, R. C. and Wilkerson, F. P.: The use of 15N to measure nitrogen uptake in eutrophic oceans, experimental conditions. Lim- www.biogeosciences.net/5/215/2008/ Biogeosciences, 5, 215–225, 2008 224 S. Bonnet et al.: The nutritional status of the South East Pacific

nol. Oceanogr., 31 673–689, 1986. Nature, 429, 292–294, 2004. Emerson, S., Quay, P., Karl, D., et al.: Experimental determina- Montoya, J. P., Voss, M., Kaehler, P., and Capone, D. G.: A simple, tion of the organic carbon flux from open-ocean surface waters, high precision tracer assay for dinitrogen fixation, App. Environ. Nature, 389, 951–954, 1997. Microb., 62, 986–993, 1996. Emery, W. J. and Meincke, J.: Global water masses: summary and Moore, J. K., Doney, S. C., Glover, D. M., Fung, I. Y.: Iron cycling review. Oceanol. Acta., 9, 383–391, 1986. and nutrient-limitation patterns in surface waters of the World Falkowski, P. G., Katz, M. E., Knoll, A. H., et al.: The Evolu- Ocean, Deep-Sea Research Part II-Topical Studies in Oceanog- tion of Modern Eukaryotic Phytoplankton., Nature, 305, doi: raphy, 49, 463–507, 2002. 0.1126/science.1095964, 354–360, 2004. Moutin, T. and Raimbault, P.: Primary production, carbon export Falkowski, P. G., Barber, R. T., and Smetacek, V.: Biogeochemi- and nutrients availability in western and eastern Mediterranean cal controls and feedbacks on primary production, Science, 281, Sea in early summer 1996, J. Marine Syst., 33–34, 273–288, 200–206, 1998. 2002. Falkowski, P. G.: Evolution of the nitrogen cycle and its influence Moutin T., Karl, D. M., Duhamel, S., Rimmelin, P., Raimbault, P., on the biological sequestration of CO2 in the Ocean, Nature, 387, Van Mooy, B. A. S., and Claustre, H.: Phosphate availability and 272–275, 1997. the ultimate control of new nitrogen input by nitrogen fixation Geider, J. G. and La Roche, J.: The role of iron in phytoplankton. in the tropical Pacific Ocean, Biogeosciences Discuss., 4, 2407– photosynthesis, and the potential for iron-limitation of primary 2440, 2007, production in the sea. Photosynth. Res., 39, 275–301, 1994. http://www.biogeosciences-discuss.net/4/2407/2007/. Guieu C., Loye-Pilot,¨ M.-D., Ridame, C., and Thomas, C.: Chemi- Obata, H., Karatani, H., and Nakayama, E.: Automated deter- cal characterization of the Saharan dust end-member; some bio- mination of iron in seawater by chelating resin concentration logical implications for the western Mediterranean, J. Geophys. and chemiluminescence detection, Anal. Chem., 5, 1524–1528, Res, 107(D15),4258, doi:10.1029/2001JD000582, 2002. 1993. Gorbunov, M. Y. and Falkowski, P. G.: Fluorescence Induction and Parkhill, J. P., Maillet, G., and Cullen, J. J.: Fluorescence-based Relaxation (FIRe) Technique and Instrumentation for Monitor- maximal quantum yield for PSII as a diagnostic of nutrient stress, ing Photosynthetic Processes and Primary Production in Aquatic J. Phycol., 17, 517–529, 2001. Ecosystems. Photosynthesis: Fundamental Aspects to Global Price, N. M.:The elemental stoichiometry and composition of an Perspectives. A. v. d. E. a. D. Bruce, Allen Press: 1029–1031, iron-limited diatom. Limnol. Oceanogr., 50, 1159–1171, 2005. 2005. Porter, K. G. and Feig, Y. S.: The use of DAPI for identifying and Holmes, M. R., Aminot, A., Kerouel, R., Hooker, B. A., and Pe- counting aquatic microflora, Limnol. Oceanogr., 25, 943–948, terson,J. B.: A simple and precise method for measuring am- 1980. monium in marine and freshwater ecosystems, Can.J.FishAquat. Raimbault P., Slawyk G., Coste B., and Fry J.: Feasibility of us- Sci., 56, 1801–1808, 1999. ing an automated colorimetric procedure for the determination Karl, D. M., Letelier, R., Tupas, L., et al.: The role of nitrogen fix- of seawater nitrate in the 0 to 100 nmol l−1 range: examples ation in biogeochemical cycling in the subtropical North Pacific from field and culture, Mar. Biol., 104, 347–351, 1990. Ocean, Nature, 388, 533–538, 1997. Raimbault, P. and Garcia, N.: Carbon and nitrogen uptake in the Karl, D. M.: Nutrient dynamics in the deep blue sea, Trends Micro- South Pacific Ocean: Evidence for efficient dinitrogen fixation biol, 10, 410–418, 2002. and regenerated production leading to large accumulation of Karl, D., Michaels, A., Bergman, B., Capone, D., Carpenter, E., dissolved organic matter in nitrogen-depleted waters, Biogeo- Letelier, R., Lipschultz, F., Paerl, H., Sigman, D., and Stal, L.: sciences Discuss., 4, 3531–3579, 2007, Dinitrogen Fixation in the World’s Oceans, Biogeochem., 57, http://www.biogeosciences-discuss.net/4/3531/2007/. 47–98, 2002. Ras J., Claustre, H., and Uitz, J.: Spatial variability of phytoplank- Katoh, H., Hagino, N., Grossman, A. R., and Ohawa, T.: Gens ton pigment distributions in the Subtropical South Pacific Ocean: essential to iron transoport in the cyaniobacterium synechocystis comparison between in situ and predicted data, Biogeosciences sp. stain PCC 6803, J. Bacteriol., 183(9), 2779–2784, 2001. Discuss., 4, 3409–3451, 2007, Kolber, Z., Prasil, O., and Falkowski, P.: Measurements of variable http://www.biogeosciences-discuss.net/4/3409/2007/. chlorophyll fluorescence using fast repetition technique I. Defin- Raven, J. A.: The iron and molybdenum use efficiencies of plant ing methodology and experimental protocols, Biochim. Biophys. growth with different energy, carbon and nitrogen source, New Acta., 1367, 88–106, 1998. Phytol., 109, 279–287, 1988. Kustka, A. B., Carpenter, E. J., Sanudo-Wilhelmy,˜ S., and Sunda, Sandstrom,¨ S., Ivanov, A. G., Park, Y., et al.: Iron stress responses in W. G.: Iron requirements for N2 and NH4+ supported growth in the cyanobacterium Synechococcus sp. PCC7942, Physiol. Plan- cultures of Trichodesmium (IMS 101): comparison with nitro- tarum, 116(2), 255-263, doi:10.1034/j.1399-3054, 2002. gen fixation rates and Fe: C ratios of field populations, Limnol. Sanudo-Wihelmy,˜ S. A., Kustka, A., B., Gobler, C. J., et al. : Phos- Oceanogr., 48, 1869–1884, 2003. phorus limitation of nitrogen fixation by Trichodesmium in the Marie, D., Partensky, F., Simon, N., et al.: Flow cytometry analysis central Atlantic Ocean, Nature, 411, 66–69, 2001. of marine picoplankton. living colors: protocols in flow cytom- Treguer´ P. and Le Corre P.: Manuel d’analyse des sels nutritifs dans etry and cell sorting, D. S. Diamond RA, NewYork, Springer- l’eau de mer. Utilisation de l’AutoAnalyser II Technicon. 2nd Verlag, 421–454, 2000. ed., Univ. Bretagne Occidentale, Laboratoire de Chimie marine, Mills, M. M., Ridame, C., Davey, M., et al.: Iron and phosphorus Brest, France, 1–110, 1975. co-limit nitrogen fixation in the eastern tropical North Atlantic, Uitz, J., Claustre, H., Morel, A., and Hooker, S. B: Vertical distri-

Biogeosciences, 5, 215–225, 2008 www.biogeosciences.net/5/215/2008/ S. Bonnet et al.: The nutritional status of the South East Paciﬁc 225

bution of phytoplankton communities in open ocean: an assess- Zehr, J. P., Waterbury, J. B., Turner, P. J., et al.: New nitrogen- ment based on surface chlorophyll, J. Geophys. Res-oceans, 111, fixing unicellular cyanobacteria discovered in the North Pacific C08005, doi:10.1029/2005JC003207, 2006. subtropical gyre, Nature, 412, 635–638, 2001. Van Wambeke F., Bonnet S., Moutin T., Raimbault P., Alarçon G., and Guieu C.: Factors limiting heterotrophic prokaryotic production in the Southern Pacific Ocean, Biogeosciences Discuss., 4, 3799–3828, 2007, http://www.biogeosciences-discuss.net/4/3799/2007/. Wagener, T., Guieu, C., Losno, R., Bonnet, S., and Ma- howald, N.: Revisiting Atmospheric dust export to the South Hemisphere Ocean, Global Biogeochemical Cycles, in press, doi:10.1029/2007GB002984, Decenber 2007.

www.biogeosciences.net/5/215/2008/ Biogeosciences, 5, 215–225, 2008